Support NeoGAF

[stupei]

The authors have verbatim called the result horrifying too. I fully agree it's a horrifying result; but people disregarding a result just because it's scary is how we ended up in this climate change mess, how people were manipulated in to believing vaccines cause autism related conditions and I'm sure I could come up with lots of other examples. I'm going to give people shit if they disregard what is currently a valid scientific result based on feelings whether they like it or not.

A result being terrifying doesn't make it untrue, and nobody has made a reasoned argument for it not being (and nobody will in this thread, because it would likely be weeks of dedicated work at minimum). The closest we've gotten is a one line comment from a research lab in glasgow, which isn't even close to enough to form a opinion counter to the paper on, but I'm sure they will be paying very close attention to the results.

Obviously this does prove that the AI is capable of learning based on a very particular subset of specific self-identified cis individuals of a specific race in a self-selected sample (by using the site in the first place) within a specific context and using that to identify visual markers. Using that to apply more broadly to the idea of the "gay" community or gay identity is over-simplification at best, and feeds into already prominent stereotypes and assumptions about queer identity at worst. Perhaps this is mostly an over-simplification in the reporting and the study itself is more nuanced in its understanding of the community at the center of its research, but the limited sample and the implied priorities that suggests make that seem somewhat unlikely.

But the argument that this research is in some way inherently beneficial to the LGBTQ community as a whole by clearly identifying and elaborating on a potential danger is honestly ridiculous and seems founded on some underlying assumption that gay people aren't already acutely aware of the dangers we face by existing. We did not need scientists to identify this terrifying reality for us and the implication that we might lack self-awareness does not make this research or its intentions inherently more noble as a result.

 

[Interframe]

Obviously this does prove that the AI is capable of learning based on a very particular subset of specific self-identified cis individuals of a specific race in a self-selected sample (by using the site in the first place) within a specific context and using that to identify visual markers. Using that to apply more broadly to the idea of the "gay" community or gay identity is over-simplification at best, and feeds into already prominent stereotypes and assumptions about queer identity at worst. Perhaps this is mostly an over-simplification in the reporting and the study itself is more nuanced in its understanding of the community at the center of its research, but the limited sample and the implied priorities that suggests make that seem somewhat unlikely.

But the argument that this research is in some way inherently beneficial to the LGBTQ community as a whole by clearly identifying and elaborating on a potential danger is honestly ridiculous and seems founded on some underlying assumption that gay people aren't already acutely aware of the dangers we face by existing. We did not need scientists to identify this terrifying reality for us and the implication that we might lack self-awareness does not make this research or its intentions inherently more noble as a result.

Pretty much sums up this study for me. It perpetuates negatives stereotypes of what a gay person is supposed to look like despite the study having a very limited selection bias and also is largely pointless considering we all know that a machine that attempts to identify sexual minorities has horrifying implications.

 

[Zoe]

But the argument that this research is in some way inherently beneficial to the LGBTQ community as a whole by clearly identifying and elaborating on a potential danger is honestly ridiculous and seems founded on some underlying assumption that gay people aren't already acutely aware of the dangers we face by existing. We did not need scientists to identify this terrifying reality for us and the implication that we might lack self-awareness does not make this research or its intentions inherently more noble as a result.

The community may know that, but that doesn't mean policy makers do. Something like this could lead to legislation over the usage of AI.

This is like white hat vs black hat.

 

[Linkstrikesback]

But the argument that this research is in some way inherently beneficial to the LGBTQ community as a whole by clearly identifying and elaborating on a potential danger is honestly ridiculous and seems founded on some underlying assumption that gay people aren't already acutely aware of the dangers we face by existing. We did not need scientists to identify this terrifying reality for us and the implication that we might lack self-awareness does not make this research or its intentions inherently more noble as a result.

Did you know concretely beforehand that technology was at the point where it could appear to be moderately successful at the automation of sexual identification through facial recognition was possible?

That's a loaded question of course, because you couldn't have known it (unless this study had already been done by others). Whether this study existed or not, I'm a hundred percent sure that governments were doing private studies for whatever nefarious purposes they would want the technology to identify people in to various groups for. The only difference is that now it's public knowledge that it's a thing that exists.

 

[stupei]

The community may know that, but that doesn't mean policy makers do. Something like this could lead to legislation over the usage of AI.

This is like white hat vs black hat.

It might lead to legislation over the usage of AI in countries that weren't already likely to use AI to murder gay people, sure.

Unless you're suggesting this would actually somehow lead to restrictions being placed on someone like Google. I guess I'm just not optimistic about that.

Did you know concretely beforehand that technology was at the point where it could appear to be moderately successful at the automation of sexual identification through facial recognition was possible?

That's a loaded question of course, because you couldn't have known it (unless this study had already been done by others). Whether this study existed or not, I'm a hundred percent sure that governments were doing private studies for whatever nefarious purposes they would want the technology to identify people in to various groups for. The only difference is that now it's public knowledge that it's a thing that exists.

Do you think the people responsible for making these technologies didn't already know the risk they're creating?

Do you think making that risk more publicly known is going to stop them?

 

[VegiHam]

The authors have verbatim called the result horrifying too. I fully agree it's a horrifying result; but people disregarding a result just because it's scary is how we ended up in this climate change mess, how people were manipulated in to believing vaccines cause autism related conditions and I'm sure I could come up with lots of other examples. I'm going to give people shit if they disregard what is currently a valid scientific result based on feelings whether they like it or not.

A result being terrifying doesn't make it untrue, and nobody has made a reasoned argument for it not being (and nobody will in this thread, because it would likely be weeks of dedicated work at minimum). The closest we've gotten is a one line comment from a research lab in glasgow, which isn't even close to enough to form a opinion counter to the paper on, but I'm sure they will be paying very close attention to the results.

That's a fair point. It is important to acknowledge it now this has has happened. I just wish they'd left well alone, you know?

The argument I'd give against the methodology is that gay dudes on dating sites may be going out of there way to look 'gayer', like ensuring eyebrows or on fleek in a way straight people aren't; and it could be biasing the input.

 

[Mr.Mike]

https://en.wikipedia.org/wiki/Full_disclosure_(computer_security)

In the field of computer security, independent researchers often discover flaws in software that can be abused to cause unintended behaviour, these flaws are called vulnerabilities. The process by which the analysis of these vulnerabilities is shared with third parties is the subject of much debate, and is referred to as the researcher's disclosure policy. Full disclosure is the practice of publishing analysis of software vulnerabilities as early as possible, making the data accessible to everyone without restriction. The primary purpose of widely disseminating information about vulnerabilities is so that potential victims are as knowledgeable as those who attack them.

This isn't strictly computer security, but I can imagine the researchers would be influenced by the other people in CS, where mostly people agree with a policy of some sort of disclosure when it comes to vulnerabilities. In this case the researchers made sure to only use publicly available tools and known methods so as to not provide any novel aid to any parties that might want to implement such a system.

 

[Lesath]

The authors have verbatim called the result horrifying too. I fully agree it's a horrifying result; but people disregarding a result just because it's scary is how we ended up in this climate change mess, how people were manipulated in to believing vaccines cause autism related conditions and I'm sure I could come up with lots of other examples. I'm going to give people shit if they disregard what is currently a valid scientific result based on feelings whether they like it or not.

A result being terrifying doesn't make it untrue, and nobody has made a reasoned argument for it not being (and nobody will in this thread, because it would likely be weeks of dedicated work at minimum). The closest we've gotten is a one line comment from a research lab in glasgow, which isn't even close to enough to form a opinion counter to the paper on, but I'm sure they will be paying very close attention to the results.

People need to keep in mind that the discovery was always there, and from what I am reading, their approach was simple enough that others would have done the same eventually.

 

[MutFox]

I want to try this...
Want to see what the AI thinks I am.

Also want to see if by changing angles and facial expressions, I can change its analysis.

 

[Eridani]

I want to try this...
Want to see what the AI thinks I am.

Also want to see if by changing angles and facial expressions, I can change its analysis.

Facial angle was detected, and the faces were properly aligned. It's something that's relatively easy to do. Facial expressions could effect it though, to an extent.

 

[valeu]

Matt Bomer has a short nose and a wide jaw. Test fail..........that, said, do we really need a scientific study for "gay face"?

 

[foodtaster]

it's basically the same as being black. IMO. the main difference between black and white is the skin color and facial features, obviously. and that's what people discriminate against. similarly, if this is able to recognize differences between gays and non-gays those who are bigoted, in how they can use their own eyes to discriminate against blacks, can discriminate against gays, too.

 

[Vermillion]

I mean it could also be that queer people might know a lot of people who don't fit these biological markers -- as well as those who do -- and are skeptical based on purely anecdotal evidence within our own lived experience within a community being studied.

Of course, and I get that. Still only anecdotal evidence.

 

[GregLombardi]

Obviously this does prove that the AI is capable of learning based on a very particular subset of specific self-identified cis individuals of a specific race in a self-selected sample (by using the site in the first place) within a specific context and using that to identify visual markers. Using that to apply more broadly to the idea of the "gay" community or gay identity is over-simplification at best, and feeds into already prominent stereotypes and assumptions about queer identity at worst. Perhaps this is mostly an over-simplification in the reporting and the study itself is more nuanced in its understanding of the community at the center of its research, but the limited sample and the implied priorities that suggests make that seem somewhat unlikely.

But the argument that this research is in some way inherently beneficial to the LGBTQ community as a whole by clearly identifying and elaborating on a potential danger is honestly ridiculous and seems founded on some underlying assumption that gay people aren't already acutely aware of the dangers we face by existing. We did not need scientists to identify this terrifying reality for us and the implication that we might lack self-awareness does not make this research or its intentions inherently more noble as a result.

A good post.

Also, I'm not a machine based AI, but I can tell you that the results of this study are not all encompassing. I still believe (specifically from a large N of personal experiences) that sexuality has layers of genetics, environment, and sometimes perhaps even in very limited cases choice that cannot be captured in this snapshot. I suspect there is much more to come on this, and not all for the good.

 

[Brawly Likes to Brawl]

Can't trust humanity with this shit.

 

[Lesath]

Matt Bomer has a short nose and a wide jaw. Test fail..........that, said, do we really need a scientific study for "gay face"?

You're sort of acting as if Matt Bomer's some sort of exception to the rule without establishing whether he falls as an exception to the algorithm in the first place. And of course, there are what is known as outliers.

A 70%-90% hit rate might seem low, but it is better than sheer random chance, and strongly points at something underlying in the biology. And if there are physiological features correlated with sexual orientation, it hints at fundamental biological genetic and environmental processes that underlie human development. Put another way, if you want to imagine a world where we understood the human biology in its entirety and eliminated disease, it seems absurd to me that such a world would happen in absence of this finding (if it is true).

 

[valeu]

we already know gay and lesbian are genetic

 

[markitrons]

The faces idea is interesting - I like that it leads credence to the idea that it's not a "choice."

The fact that it provides evidence to back up the idea of hormone exposure is scary to me. Seems like too easy of a leap to spin being gay as a negative to be eradicated in pregnancy via genetic manipulation or something similar.

 

[MogCakes]

The fact that it provides evidence to back up the idea of hormone exposure is scary to me. Seems like too easy of a leap to spin being gay as a negative to be eradicated in pregnancy via genetic manipulation or something similar.

It is already inevitable that people will try that angle once the gay as a choice rhetoric dies. No reason to fear if it is already guaranteed to happen. Hate is hate.

 

[valeu]

to me it feels like an attack on gay and lesbians. nobody wants their sexuality or features catagorized like this.....if scientists make a study that gives a percentage number of how maculine or feminine heterosexuals are based on their face, voice, and body language, you will have a lot of furious heteros

 

[Battersea Power Station]

to me it feels like an attack on gay and lesbians. nobody wants their sexuality or features catagorized like this.....if scientists make a study that gives a percentage number of how maculine or feminine heterosexuals are based on their face, voice, and body language, you will have a lot of furious heteros

Is there any data to back up any of your claims? Just anecdotally, I know quite a few people personally who actively try to project their sexuality through either fashion or by accentuating certain physical features.

 

[SomedayTheFire]

Sounds like bullshit. Utterly terrifying and very very dangerous.

Is there any data to back up any of your claims? Just anecdotally, I know quite a few people personally who actively try to project their sexuality through either fashion or by accentuating certain physical features.

By choice, not force.

 

[StanleyStutters]

there are two ways this can go.

1. Its more proof that people are born a certain sexuality and that it isn't through choice and may further foster acceptability of LGBT's

2. it can be used to identify and persecute "the gays" and may lead to modern day witch hunts if used incorrectly

If it turns out to be legit, which is still a ways off, I fear option 2 is the more likely scenario in the world we currently live in.

 

[kirblar]

Huh. Never actually thought about identical twins. Didn't know that identical twins could have different sexualities. Would have thought they were identical in that respect too.

Chances are actually higher that one will be gay/Trans w/ a set of twins relative to an individual child.

 

[valeu]

Is there any data to back up any of your claims? Just anecdotally, I know quite a few people personally who actively try to project their sexuality through either fashion or by accentuating certain physical features.

how many men that you have met in your life, gay or hetero, like being called feminine? a sissy? this is the go-to insult of bullys

 

[Linkstrikesback]

Do you think the people responsible for making these technologies didn't already know the risk they're creating?

Do you think making that risk more publicly known is going to stop them?

That's just it: The people doing this study didn't make any new technology. The facial recognition software they used, FACE++ is already out there, and anyone can buy time using it. And frankly, measuring curvatures and sizes of objects in images isn't exactly novel either.

As I said already, the only way this would go away is if all facial recognition technology was banned, and that ship has long sailed.

As for publicly stopping them? I wish. If I knew any way to efficiently affect the policies of governments, I would be rich. But I on a personal level would like to know as much as possible concretely about the things I suspect people are up to behind the scenes. This obviously is not going to console at all anyone in any of the countries that have punishments for being anything other than straight in to the laws of the country they reside in. But this result was always coming. If you really believe being gay is not a choice, which I hope most gaffers do! then it's hardly surprising that there would be other biological signs, including ones that could eventually be used to distinguish people in to groups by these categories.

 

[Battersea Power Station]

how many men that you have met in your life, gay or hetero, like being called feminine? a sissy? this is the go-to insult of bullys

I know 3, but that's not relevant. Is the claim here that this algorithm will embolden bullies in calling others gay? I think that's specious at best and anti-science at worst.

It's akin to saying Project Implicit is responsible for racial discrimination.

You don't deal with bias by plugging your ears and pretending it doesn't exist.

 

[stupei]

Here's the thing: if their actual concern is with the dangers of this technology and exposing or demonstrating how it could be used negatively, I'm not sure why they felt they had to use an already vulnerable minority group as their sample when attempting to identify what they believe to be clear visual markers that could be replicated and used to identify members of said group without their consent.

But of course a research paper that claims to be able to identify sexual orientation is going to get a lot of headlines and eyeballs as compared to other categories one might use to demonstrate the accuracy of facial recognition technology.

It feels wildly irresponsible, no matter what they might suggest their intentions are.

Maybe I'm missing some context, but what other form of the study could they have done?

If you're a sociologist studying the nature of sexuality and sexual preference and the intersection of genetics and physical appearance, isn't this data invaluable?

If your premise is "in the wrong hands, this kind of technology could kill people," then you could use literally anything else to prove that the technology is advanced enough to identify subtle biological factors that the human eye cannot perceive that are not inextricably linked to the persecution of entire categories of people.

If you are doing this because you study sex and sexuality, then you're not actually doing it as a warning or because you believe it is dangerous and don't have that moral ground to pretend to stand on in the first place.

And frankly I'm not especially thrilled with the idea that someone's area of expertise would actually be the intersection of genetics, physical appearance, and sexual orientation. I don't think that it's anti-science to suggest that some areas of interest border dangerously close to foundational beliefs that might easily lead to forms of eugenics.

 

[Battersea Power Station]

It feels wildly irresponsible, no matter what they might suggest their intentions are.

Maybe I'm missing some context, but what other form of the study could they have done?

If you're a sociologist studying the nature of sexuality and sexual preference and the intersection of genetics and physical appearance, isn't this data invaluable?

 

[Alcoremortis]

Isn't this the AI that was compiled using existing AI with new search parameters?

If that's the case, I'm glad they did the study. The tools are already out there, if these scientists hadn't done it, someone else, possibly someone more nefarious, would have done it first. At least here if it's legit, everyone will know this tool can be used against them and possibly be able to take precautions. And if it's fake, it will be publicly discredited and therefore hinder other attempts to use a similar tool for evil means.

 

[MsKrisp]

Eh, I can't help but question the methodology here. Collecting photos from dating sites along with reported sexualities was the cheapest and fastest way to collect a lot of data to use in order to make this possible. But photos for dating sites are taken from many different angles, and photos could be altered or the person could have had plastic surgery. Without consistently measuring or using photos taken in a consistent manner, you can be sure how accurate this AI really is.

 

[SomTervo]

Whoa. Now THIS is interesting.

 

[m0dus]

Well that much is obvious. They can't make up their minds—is it "junk science" that doesn't work, or something that legitimately threatens to out people who aren't ready to be outed?

Until a peer-reviewed paper is out to be scrutinized it's all just whistling dixie.

Because it's both junk science and dangerous, something doesn't have to be legitimate to be misused, just accepted by people who don't understand it.

For example: the polygraph.

 

[Linkstrikesback]

Because it's both junk science

I'm not going to argue the it's dangerous part, it obviously has the potential to be horribly misused.

However, why do people keep calling it "junk" science? Am I misunderstanding what people mean by that?

Because accusing them of publishing fraudulent results (which is what junk science typically means whenever I've heard it used by other scientists, e.g. about Andrew wakefield) is basically the biggest insult to any scientist if you're wrong about it. And the preprint has been published for less than a week, so I'm still fairly sure that no-one has done any work disproving the result yet.

 

[Terrell]

I still don't understand why it isn't a positive overall by proving people are born gay and it's not a choice or something that can be burned out of a person somehow.

All it will take is one person deciding that it's an undesirable genetic trait and use this app to perform a pogrom on gay people. One person.

Yah agree. The biggest slippery slope right now is anti science and while I get their concern on usage there's been too much politics in science already. If it's wrong it needs to be proven to be wrong scientifically.

Just like climate change, right? People who don't believe climate change is real do so purely based on facts, right?

When you apply that same human pattern of behaviour regarding science and believing facts that aren't peer-reviewed towards identifying a marginalized and sometimes-hated segment of the population, you put the lives of those marginalized people in physical or mortal danger.

You can be pro-science and still agree that, until such a time that society doesn't actively kill gay people, saying there is technology that could identify them as such against their will is dangerous, even if it's scientifically debunked. But feel free to debate about my safety by saying that the scientific community saying something is wrong is going to amount to a hill of beans.

 

[CornBurrito]

I still don't understand why it isn't a positive overall by proving people are born gay and it's not a choice or something that can be burned out of a person somehow.

People being born gay doesn't require them to have gay sex. Proving that people are born gay won't stop oppression.

Anyway, I dislike the trend here of scientific progress without concern for ethics. These scientists developing facial recognition/gay recognition AIs have to be ignorant if they think these won't be used as tools of oppression. Imagine trying to protest against the government when they can put your face in a database and then use it to track your location 24/7.

 

[Majmun]

Hmm interesting. I hope it can clear my doubts about Frankie Grande.

 

[You Are Viewtiful]

Hmm interesting. I hope it can clear my doubts about Frankie Grande.

Isn't he openly gay?

 

[Linkstrikesback]

All it will take is one person deciding that it's an undesirable genetic trait and use this app to perform a pogrom on gay people. One person.



Just like climate change, right? People who don't believe climate change is real do so purely based on facts, right?

When you apply that same human pattern of behaviour regarding science and believing facts that aren't peer-reviewed towards identifying a marginalized and sometimes-hated segment of the population, you put the lives of those marginalized people in physical or mortal danger.

You can be pro-science and still agree that, until such a time that society doesn't actively kill gay people, saying there is technology that could identify them as such against their will is dangerous, even if it's scientifically debunked. But feel free to debate about my safety.

This has been peer reviewed, which is a requirement for publishing in any decent journal (including the one this was submitted to). It will get further peer review now it's completely in the open, but scientists don't make it a habit to keep research a secret forever. And even if they didn't publish the article, the technology wouldn't stop existing. They've not invented anything new in the study. I'm very much sure private companies and governments have been making programs that can do similar analysis, except they don't have to make it public knowledge. The only difference between the two possible situations is whether the general public is aware of it existing or not.

You can make a very good argument that the world would be a better place without it existing, but without going back and preventing facial recognition software from ever being invented, that's not an option.

 

[CornBurrito]

Honestly it's fucking funny that we have posters here going "well if we can prove homosexuality is genetic, homophobia will be eradicated!"

How much melanin you have in your skin is genetic, and most of the developed world understands that even if they never picked up a biology textbook. So answer me this: is racism gone?

 

[Terrell]

This has been peer reviewed, which is a requirement for publishing in any decent journal (including the one this was submitted to). It will get further peer review now it's completely in the open, but scientists don't make it a habit to keep research a secret forever. And even if they didn't publish the article, the technology wouldn't stop existing. They've not invented anything new in the study. I'm very much sure private companies and governments have been making programs that can do similar analysis, except they don't have to make it public knowledge. The only difference between the two possible situations is whether the general public is aware of it existing or not.

You can make a very good argument that the world would be a better place without it existing, but without going back and preventing facial recognition software from ever being invented, that's not an option.

Facial recognition is fine. The algorithm for identifying gay people wasn't necessary and isn't intrinsic in said technology. One could exist without the other. All it takes is self-restraint and the knowledge of what a damaging idea it is.

And yeah, if there's work going on to the same effect that isn't being published, I'd equally have concerns about that, it's not like these scientists are the sole focus of my concern, the study of such a thing itself is.

But by publishing that information, it gives others who might have hit a dead end on that something to galvanize their effort. It makes it plausible and publicized for the entire world, which makes it all the more dangerous.

 

[Lesath]

Facial recognition is fine. The algorithm for identifying gay people wasn't necessary and isn't intrinsic in said technology. One could exist without the other.

And yeah, if there's work going on to the same effect that isn't being published, I'd equally have concerns about that, it's not like these scientists are the sole focus of my concern, the study of such a thing itself is.

But by publishing that information, it gives others who might have hit a dead end on that something to galvanize their effort. It makes it plausible and publicized for the entire world, which makes it all the more dangerous.

Dead end? Dude, this is low-hanging fruit. Literally an observational study.

 

[Linkstrikesback]

But by publishing that information, it gives others who might have hit a dead end on that something to galvanize their effort. It makes it plausible and publicized for the entire world, which makes it all the more dangerous.

Anyone who hit a dead end and didn't think of the methodology they used which is essentially measuring the curvatures of various facial features isn't going to be able to replicate this even with this published. It's a very well established technique already that has existed for a long time in various microscopy studies.

 

[Terrell]

Anyone who hit a dead end and didn't think of the methodology they used which is essentially measuring the curvatures of various facial features isn't going to be able to replicate this even with this published. It's a very well established technique already that has existed for a long time in various microscopy studies.

And to people who hadn't considered the possibility that it could be done or consider using that "established technique" since it's one of several and/or their particular logistic regression algorithm which would have to have been tailored to this task, there's now a race to see who can make it 100% accurate first.

 

[michaelius]

Obviously this does prove that the AI is capable of learning based on a very particular subset of specific self-identified cis individuals of a specific race in a self-selected sample (by using the site in the first place) within a specific context and using that to identify visual markers. Using that to apply more broadly to the idea of the "gay" community or gay identity is over-simplification at best, and feeds into already prominent stereotypes and assumptions about queer identity at worst. Perhaps this is mostly an over-simplification in the reporting and the study itself is more nuanced in its understanding of the community at the center of its research, but the limited sample and the implied priorities that suggests make that seem somewhat unlikely.

That's how teaching perceptron neural networks works (mind you I don't know what they use) . You give it a learning set which it uses to set the weights which it uses to calculate result.
While the set they are using might be not representative it at least work partially and if such network is feed more accurate data results will improve.

 

[Massive Duck. C.M.]

I don't really understand what positive applications there are for such a thing.

 

[Cybit]

Because it's both junk science and dangerous, something doesn't have to be legitimate to be misused, just accepted by people who don't understand it.

For example: the polygraph.

It has a long way to go before it can be called 'junk science' - it will go through the peer review process, folks will try to replicate it with bigger sets of data, and then folks will figure out whether it's legitimate or not.

Our personal feelings on the morality of the progress is not the condition as to determine whether it is factual or not. That's the same argument climate change deniers use.

People being born gay doesn't require them to have gay sex. Proving that people are born gay won't stop oppression.

Anyway, I dislike the trend here of scientific progress without concern for ethics. These scientists developing facial recognition/gay recognition AIs have to be ignorant if they think these won't be used as tools of oppression. Imagine trying to protest against the government when they can put your face in a database and then use it to track your location 24/7.

Here's the thing that people seem to be missing (and it is obvious that almost everyone freaking out about the ethics of such things haven't actually read the article, since the authors DO THE SAME THING) - this was not something they "developed". This is something they used publicly available software and methods to do. This was already sitting there to be used, except it would have been done so while we are unaware of those capabilities. If you go to the Super Bowl or any major event, facial recognition software and hardware that is years beyond what the public domain has is used on you, for instance. If the Snowden leaks didn't illuminate the level of "shit that the government is working on that is years ahead of anything public", I don't know what else could. Apple, Google, Microsoft, and other software and hardware manufacturers had no idea that the NSA had found backdoors into all of their systems (or negotiated them), and had Snowden not done what he did, all of those companies would probably be blissfully unaware that their data was not secure. That knowledge let them patch those holes.

The idea that "oh no, now people know it is possible!" is incredibly ridiculous and ignorant of history. If you're a government or group who is already motivated to try to use this for nefarious purposes, you've already been working on this and have gotten much further than a pair of random folks using freaking online dating pictures. Heck, a thorough study of this might find weaknesses in the algorithm that allow those who are rightfully worried about misuse to counter it. But if you don't study it, you won't ever get there, and you just have to pray that no one else decides to see if it is possible.

 

[NHarmonic.]

this

I respect the science, but I can easily see the dangers of such tech if public opinion shifts, oppressive governments want to mass target homosexuals, or even smaller groups of people who want to make the lives of people a living hell.

Think about what would happen if this technology gets to ISIS. They've already killed a lot of LGBTQ+ folk, with this they are putting a lot of people in danger.

 

[Enter The 36 Chambers]

How long before snapchat gets this tech.

Definitely gonna be a new filter by 2019